brief academic opinion of economic professors and scholars ... filebrief academic opinion of...

Trends in German and European Electricity Working Papers

WP-GE-09

Brief Academic Opinion of Economic Professors and Scholars on the Project of

Acquisition of Endesa by Gas Natural Julian Barquin, Pontificia Comillas University (Spain) Lars Bergman, Stockholm School of Economics (Sweden) Claude Crampes, University of Toulouse (France) Jean-Michel Glachant, University Paris XI (France) Richard Green, University of Birmingham (United Kingdom) Christian Von Hirschhausen, Dresden University of Technology (Germany) François Lévêque, Ecole des mines de Paris (France) Steve Stoft (Berkeley, USA)

Dresden University of Technology Chair of Energy Economics and

Public Sector Management

Brief academic opinion of economic professors and scholars on the project of

acquisition of Endesa by Gas Natural

October 26, 2005

Julian Barquin, Pontificia Comillas University (Spain)

Lars Bergman, Stockholm School of Economics (Sweden)

Claude Crampes, University of Toulouse (France)

Jean-Michel Glachant, University Paris XI (France)

Richard Green, University of Birmingham (United Kingdom)

Christian Von Hirschhausen, Dresden University of Technology (Germany)

François Lévêque, Ecole des mines de Paris (France)

Steve Stoft (Berkeley, USA)

Brief academic opinion of economic professors and scholars on the acquisition of Endesa by GN

2

Table of Contents

Executive summary ___________________________________________________________________ i

Introduction _________________________________________________________________________3

1. The adverse effects of the transaction on the building of the European energy market_____________7

The erection of dual fuel barriers to entry as an obstacle to the European internal energy market ____8

Fewer incentives for the Europeanisation of energy companies _______________________________9

A likely decrease in cooperation to achieve the construction of the European internal market ______10

2. The economic rationale for antitrust authorities to be very cautious in clearing transactions related tothe electricity sector _________________________________________________________________12

Mergers in the electricity sector raise the concern of type II errors rather than type I errors_________12

Role and drawbacks of remedies ______________________________________________________13

3. Anticompetitive effects of gas-electricity mergers ________________________________________15

Four main anticompetitive effects _____________________________________________________15

The case of the projected GN/Endesa merger_____________________________________________17

Horizontal effect on the wholesale electricity market ____________________________________18

Horizontal effect on retail markets___________________________________________________19

Vertical effect on the wholesale electricity market ______________________________________20

Vertical effect on the wholesale gas market____________________________________________22

References_________________________________________________________________________23

Appendix: Short bios of signatories to the brief _______________________________________________24


3

Executive summary

This opinion describes what the signatories believe are consensus views in academia onthree major economic questions that arise with the projected acquisition of Endesa by GasNatural.

1. Do national gas-electricity mergers strengthen or deter the building of the European

internal energy market?

The belief that national energy champions would make competition more vibrant at theEU level is wrong. On the contrary, the creation of giants combining gas and electricitywithin national boundaries is likely to slow down the building of the internal market for thefollowing reasons:

- National gas-electricity mergers create dual fuel barriers to entry that limit possibilitiesof foreign producers trying to gain a foothold in another EU market while contestableentry to electricity, helped by gas liberalisation, should be an alternative option t oenlarge the competitive arena before sufficient new interconnections are built.

- Gas-electricity national champions have less incentive to grow beyond their borders.The proposed divestitures of Endesa’s existing activities in other EU countries may justsignal the national tropism of the new entity.

- The clearance of a locally anticompetitive merger generates an external effect byincreasing suspicion of the Member State’s fair play in building the internal energymarket, while cooperation between Member States is crucial for achieving the Europeaninternal energy market.

2. For the sake of electricity consumers, should ineffective and overly-stringent remedies be

viewed as equally undesirable?

In order to protect consumers, economic theory recommends that merger control inelectricity markets should be more cautious and stringent than in other sectors. That is, theantitrust authorities should be more willing to risk imposing overly-stringent pro-competitive remedies (type I errors) than to risk prescribing ineffective remedies (type IIerrors).

Due to the extremely weak demand elasticity, electricity markets are extremely susceptibleto market power. Consequently, a type II error can easily cause a huge transfer of surplusfrom consumers to producers and force final consumers pay nearly the full cost of theanticompetitive outcomes of inappropriate mergers. By contrast, the cost of prohibiting


4

mergers that provide efficiency gains (type I errors) will typically do very little harmconsumers and suppliers, because significant efficiency gains from mergers are difficult t oachieve in the electricity and gas sectors.

To keep the good (efficiency gains) without the bad (increased market power), competitionlaw empowers antitrust authorities to impose merger remedies. But in the electricity and gassector, the use of merger remedies is a risky game. Remedies can fail in a number of ways.Moreover, the risk of error is higher in markets recently opened to competition becausecompetition authorities have to make guesses not only on the effects of the merger and onthe effects of the remedies, but also on the future mode of competition and marketboundaries before they are stabilised.

3. What are the potential anticompetitive effects of the projected acquisition?

Removal of an effective and growing independent competitor

Because of some divestments by Endesa to Iberdrola, the Gas Natural/Endesa transaction isdesigned so that the total generation capacity of the merged company would be below thecurrent capacity of Endesa. However, such a lowering may hide a market power increase inthe energy market:

1. The market power of generators depends on whether for a given time (especially inpeak hours) and a given place (in load pockets) a generator is indispensable to servedemand and to balance the system. Therefore, the unilateral effect of the mergergreatly depends on the redistribution of plants.

2. As market power exercised by others has the same effects on price without the costs ofwithholding, the merged company can profit just as much by giving market power to anexisting competitor as to itself.

3. The proposed merger could increase the risk of collusion in the Spanish wholesaleelectricity market, as it would remove the pressure of a strong independent new entrant.The bargaining over remedies between incumbents could also increase the symmetrybetween the two major Spanish producers and facilitate collusive practices.

Elimination of strong and effective dual fuel competitors

Where both Gas Natural and Endesa are incumbent distributors and retailers, the mergerwould eliminate a key dual fuel competition. This is because the two incumbents havestrong local customer bases, established reputations and can save money in bundling gas andelectricity. Moreover it would instantly give to the merged company the advantage ofbecoming a dual fuel incumbent.


5

Input foreclosure

Whereas the other anticompetitive effects would simply terminate one source ofcompetition, the vertical integration of GN and Endesa would confer additional marketpower on the merged company if it was able to profitably raise the price of gas charged t ocompeting generators. The merger could increase the market power of the new entity inthree different ways that are detailed in the opinion.

Customer foreclosure

In substituting exchanges through the market by exchanges within a single integrated firm,the merger may impede the development of the wholesale gas market, a key aspect of gasliberalisation. The elimination of a key purchaser may also decrease economies of scalethat alternative gas suppliers would have achieved absent the merger, and thus may result ina price increase for their sales. If the integrated firm has enough market power on thewholesale gas market, it will follow the price increase rather than offer a lower price. In theshort run, the foreclosure effect may concern only a small amount of gas (e.g. in excess ofvolumes contracted in the long term), but this would still contradict the objective of anintegrated competitive gas market, that is supposed to achieve shorter contract durationsand more liquid wholesale markets.

To sum up, vertical and horizontal anticompetitive effects of convergent mergers areseveral, and they may be individually strong. Clearing the merger with ineffective remediescould lead to tremendous costs born by final consumers. The merger would increase dual fuelbarriers to entry in Spain, reduce the incentives for the merged company to grow outside itsnational borders, slow the building of the internal market, and reduce regional cooperationin Europe. We therefore strongly recommend that the pro- and anti-competitive featuresof the proposed GN/Endesa merger be accurately and thoroughly assessed using a detailedmodel of the electricity market according to established antitrust standards.


6

Introduction

The signatories of this opinion have researched and written extensively on power systemeconomics and competition policy (see appendix). This brief describes what we believe areconsensus views in academia on some economic questions that arise with the acquisition ofEndesa by Gas Natural.

In section 1, we discuss the creation of national giants in gas and electricity from aEuropean energy internal market perspective. Our concern is that the transaction may slowdown the building of the internal market. The following two sections adopt an antitrustperspective. Section 2 explains why antitrust authorities must be very cautious aboutclearing mergers in the electricity sector. Section 3 provides an overview of the potentialvertical and horizontal effects of gas-electricity mergers.

The opinion results from research based on publicly available information. The researchwas financed by Endesa with the funds contributed to non profit research organisations andfoundations. The opinion, however, only reflects the views of the signatories.


7

1. The adverse effects of the transaction on the building of the European energy market

It may seem puzzling to start with this concern. European energy law ignores mergers andacquisitions (hereafter M&As) and, for better or for worse, it does not empower a FERC-like institution to have a say in merger control. Competition law enforcement is the onlyEU standard. Moreover, whereas economic scholars are familiar with the assessment of theanticompetitive effects of a merger on a relevant market, they are far less well equipped t oanalyse market concentration from a market-development perspective.

However, one false belief and one salient fact have led us to comment on this concern.

Conventional wisdom holds that national energy champions would make competition morevibrant at the EU level. We agree that in football we would like to see a merged RealMadrid/Barça fighting with a merged Chelsea/Arsenal. We regret having to disagree with thisopinion when it is applied to European gas and electricity industry, but economiccompetition works by different principles than does competition in sports. In sportstournaments the ultimate competition is between the top two contestants. In economicsperfect competition is mainly approached as the number of competitors becomes verylarge, while competition between two giants is a duopoly. We have not seen yet anyorganisation of a yearly European energy championship, and are not aware that energycompetitors have to put at risk the cups they won in previous seasons of the tournament.Furthermore, we do know that access to the industry playing field is open only whenbarriers to trade are removed, and that the easiest way to win a local industry race is t oclose access to the track lanes and the stadium to anyone who is not a member of the samenational club. Moreover, if all European electricity and gas champions barricade access t otheir own stadiums when, where and with whom they will start training for the Europeancup?

The fact is that from 1998 to 2003, there have been about 135 M&As in the EU gas andelectricity sector with about one third being cross-border and two thirds national (Codognetet alii, 2003).

We would be surprised if this flow of M&As had no link with the design of the internalmarket and no effect on it. As scholars we are curious to know if this effect is positive ornegative, because either is possible.


8

Number of mergers of European electricity companies from 1998 to 2003 (Source: Codognet et alii, 2003)

In the remainder of this section, we propose to analyse the impact on the construction ofthe European internal market of the acquisition of Endesa by Gas Natural (Hereafter, GN).We consider three types of effects: the contribution of the Spanish market to the Europeaninternal market; the Europeanisation of the energy companies operating in the differentMembers States; and the cooperation between national institutions (e.g., regulatoryauthorities) in the building of the European internal market.

The erection of dual fuel barriers to entry as an obstacle

to the European internal energy market

There are two routes to effective competition in electricity and gas markets: the directroute is to ensure that production capacities are divided between sufficiently manycompeting companies so that no one has much influence over the price; the indirect routeis to expose incumbent operators to a credible threat of entry if the price rises above thecompetitive level. But given the important and still growing concentration of both gas andelectricity industries in many EU countries and given the constrained cross-borderinterconnections, the last door open for an electricity company to enter into a foreignmarket is often to generate electricity with gas bought from an “open-door to electricity”national gas company who could, in addition, offer its forces in the retail market to sell dual


9

fuel services. Reciprocally, a foreign gas supplier could expect to enter the gas market byselling gas to an “open door to gas” national electricity generator, with a possiblecooperation in the retail market to sell dual fuel offers. Furthermore, even relatively smallentries in gas and electricity could encourage each other because the size of an efficient gasfired plant can be relatively small1. Lastly, contestable entry and gas liberalisation reducetransmission constraints (see D. Newbery, 2005). They are an alternative option to enlargethe competitive arena whenever investments in interconnections are hindered by local ‘Notin my backyard’ syndrome and the national tropism of TSOs 2.

In contrast, the GN/Endesa transaction would render the Spanish industry structure morevertically compact and subsequently more closed to foreign initiative. Electricitygeneration, being the biggest and fastest growing customer of the gas industry, would bevertically embedded in the gas industry. The leading electricity and gas incumbentcompanies would instantly become dual fuel operators before any other existing orpotential competitor has a chance to do so. Then, the barriers to enter the market fromabroad would be higher and the risk taken by entrants would be higher too. If such anindustrial structure was to be generalized in the EU, the dual fuel barriers to entry wouldmake the corresponding markets much tougher for outsiders. This would dramatically affectthe growth possibilities for the competitive fringe (being made of small new entries incountries’ incumbent yards). As a result, the European internal market would be madesmaller and not bigger, with adverse effects for all European consumers. It would be reducedto very small spots and thin shores. One must keep in mind that the size of the internalmarket is not equal to the sum of all national markets but only to the parts that are trulyopen to competition in each national market.

Fewer incentives for the Europeanisation of energy companies

Once each country is dominated by a national giant that integrates gas and electricity, itwill be rare to find these champions engaging in cross-border competition. From thisperspective, it is important not to fail to differentiate European energy firms competing onseveral regional and national markets without being dominant, and national championsdominating their national markets.

The reasons for lower incentives are threefold:

Firstly, their incentives to compete against each other are lower. With all countries havinghigher barriers to entry and tougher markets for entrants, why would a reasonable managerof any national gas-electricity champion spend time and money and take entry risks t o

1 For example, the German and Swiss T-Power is going to build in Belgian Flanders a 400 MW plant; itcould purchase gas from Norwegians or from Germans thanks to the gas hub of Zeebrugge.2 In the medium term, the focus should be kept on the expansion of transmission capacities, both for


10

challenge another giant in its closed foreign territory? If the purpose was to make money,would profit expectations be worth the risk? If the purpose was to invest for the futuregrowth of the company, while there is already plenty of investments to coordinate fromupstream gas to downstream electricity to serve a growing market at home, would it be thatattractive to invest abroad in more risky affairs? For an answer, we may have a look at theworld’s biggest gas and electricity M&A ever made between two national champions,namely Eon/Ruhrgas in 2002. Since the merger, how many foreign moves has Eon decidedto invade any other national champion’s territory? Actually none 3.

Secondly, the bigger players are not necessarily the more agile in playing outside their homemarkets. It is true that size matters in order to increase economies of scale and scope.Moreover, an overly small company cannot expect to play an important role on theEuropean scene because investigating and managing risky foreign scenarios consume time,money, people and skills that are all scarce resources. However, there are big and big. Untilrecently German RWE and Italian ENEL, each several times bigger than Vattenfall, had lessinvested than the Swedish company in foreign energy business4. It is not so obvious thatSpanish players are too small at present. For instance, in the larger US electricity market,very few electricity companies can exhibit a bigger size than Endesa.

Furthermore, following a merger between GN and Endesa, much of these companies’ timeand skill will be concentrated on making the acquisition works. As a result, there is a seriousconcern that the resources of the merged company available to acquire a truly Europeandimension would not grow after the merger, but could rather decline for some years. Theproposed divestitures to Iberdrola of Endesa’s existing activities in other EU countries maymerely signal the national tropism of the merged company.

A likely decrease in cooperation to achieve the construction

of the European internal market

Our concern is that the clearance of anticompetitive mergers may decrease the willingnessand the ability to cooperate in the making of the European internal energy market.

The current EU legal framework for building the internal energy market is not elaboratedenough to ensure that the goal will be achieved. The existing energy European directivesand regulations have themselves been designed to give room to many countries in their

electricity (Spain-France, Spain-Portugal) and natural gas (import pipelines, LNG terminal). 3 Eon’s investments in UK are precisely targeted in a market with a rather low concentration levelcompared to EU standards.4 Vattenfall is an interesting case to look at. It has quite the same home market share as Endesa in Spain(roughly half the market), but the Swedish electricity market is actually smaller that the Spanish andSwedes barely consume gas. All this did not stop Vattenfall from entering half a dozen foreign countries,while it was unable to deter in Sweden the entries of the Finnish operator Fortum (through aparticipation in Stockholm Energy) and of the German operator E.ON (through a participation in


11

national implementations. Moreover, no common European regulatory agency has beencreated to fill in the gaps, for instance, in order to coordinate the required actions t oenhance congested interconnections and favour cross-border trades. Legal compliance istherefore not the alpha and omega of the internal market.

Cooperation, especially at regional level, has appeared as a necessary supplementaryingredient. As recently, pointed out by a group of European scholars (J.-M. Glachant andF. Lévêque, 2005), the future of the internal energy market crucially depends on collectiveactions between national regulatory authorities, TSOs and power exchanges.

Of course, cooperation is confronted to free ridding and trust has to be built and maintained.In this area also, legal compliance is not the alpha and omega. In fact, legal compliancewith EU energy law is not sufficient to induce cooperation. For instance, the Frenchperfectly legal situation where a 100% State-owned enterprise enjoys a share of about 95 %in a 450 TWh national market open only to the minimal EU Directive requirement hasraised conflicts between France and other Member States. Neither is legal compliance withnational merger control law. The legal clearance of the transaction between Eon andRuhrgas by the German government is a good example. The merger has changed thearchitecture of German interests both at home and abroad. It has increased suspicion aboutGermany’s fair play in contributing to the energy internal market.

In other terms, the clearance of an anticompetitive merger may signal a national defectionin cooperating in the EU game for the building of the European internal energy market. Bycontrast, a rigorous assessing of mergers and their prohibition whenever they areanticompetitive reinforce trust and strengthen future cooperation.

We therefore strongly advocate that the GN/Endesa transaction be accurately and deeplyassessed on its own pro and anticompetitive features according to the well-recognised andestablished antitrust standards.

Sydkraft).


12

2. The economic rationale for antitrust authorities to be very cautious in clearing

transactions related to the electricity sector

According to a widely accepted legal and economic standard, mergers and acquisitions canonly be authorized if consumers’ welfare does not decrease as a result of the transaction.The economic assessment consists in comparing the efficiency gains from the transactionthat will be passed on to consumers with the anticompetitive effects of the transaction. Ifthe latter are larger than the former, the transaction should be prohibited. However, becausecompetition authorities face informational constraints, errors are inevitable. Certain pro-competitive transactions are prohibited, so-called type I errors, whereas certainanticompetitive transactions are approved, so-called type II errors. This section discussesthe effects of errors in the case of mergers related to the electricity sector, that is mergersbetween an electricity and a gas company (i.e., convergent mergers) and between twoelectricity companies (i.e., non convergent mergers).

Mergers in the electricity sector raise the concern of type II errors rather than type I errors

As it is well known, electricity generation features several fundamental characteristics thatgive the industry a high potential for market power (see S. Stoft, 2002). Buyers andmarketers in the middle cannot store the product to defend against a sudden price increase.Most end-users pay prices that are averaged over time and therefore are not responsive t oshort term price fluctuations (i.e., demand is inelastic). Both production and transmissionface capacity constraints: i) the marginal cost of production at a power station steeplyincreases when it operates near or at its capacity limit - thus supply is also inelastic at peaktimes; ii) the transmission network may experience congestion - then the relevantgeographical electricity market narrows and a single or a few producers may dominate theisolated local market. In a nutshell, traditional constraints on price increases, namelydemand response in the relevant time frame, consumers’ defection, supply substitution andpotential entry are naturally weak in electricity markets.

It is also well known that in electricity the exercise of market power can be very damagingfor consumers’ welfare. Due to weak demand elasticity, a seeming small reduction in thecompetitiveness of the market can result in a huge transfer of surplus from consumers t oproducers. Protecting consumers from market power in electricity markets is thereforeessential and type II errors can be tremendously costly for them.

Of course, type I errors are also costly for consumers. They prevent efficiency gains thatcould be passed on through lower prices. However, in electricity the costs of type I errorsshould not be overestimated. Efficiency gains that may result from mergers are difficult t oachieve. Because of unanticipated events and difficulties in integrating activities, expected


13

cost savings do not frequently materialise. According to Anderson (1999), only 15 % ofM&As among US electric utilities have achieved the financial objectives that wereexpected. This pattern has been confirmed recently by Becker-Blease et al. (2004) on asample of 152 proposed convergent and non convergent M&As between 1990 and 2002.

Role and drawbacks of remedies

Competition law empowers antitrust authorities to clear mergers with conditions. If aremedy eliminates the anticompetitive effect raised by a transaction, a clearance undercondition is a way to get the good (i.e., efficiency gains) without the bad (i.e.,anticompetitive effects), whereas a merger control based on a pure ‘yes’ or ‘no’ decisionwould lead to a welfare decreasing clearance or prohibition.

Yet, the possibility of clearance with conditions leads to new types of mistakes. Weak typeI errors occur whenever a merger is cleared with unnecessary conditions and weak type IIerrors happen whenever a merger is cleared with remedies that are ineffective in eliminatingthe anticompetitive effects.

The latter errors happen because remedies can go wrong in a number of respects:

Firstly, remedies can go wrong due to a combination of informational asymmetries andincentives of the parties that are not in line with the objective of restoring competition(M. Motta, 2004). For instance, the merging parties have better information on the assetsthey propose to divest than potential buyers and competition authorities. They can designa package of remedies that does not allow the buyer to be a viable competitor. Moreover, amerging firm’s interest is not to select an aggressive competitor as a buyer. It is alsoobvious that an aggressive competitor is not always the most able to buy the divested assets,as he may not expect profits from the acquisition as high as a rival bidder with a softerstrategy.

Secondly, remedies may facilitate collusion, especially when the divested assets increasesymmetry and multi-market contacts (M. Motta and al., 2003). The risk of single-firmdominance decreases, as a competitor is made more powerful. However, to the extent thatremedies redistribute capacity, market shares and other assets in a more symmetrical way,the risk of a collusive outcome increases. When the buyer is active in a neighbouringproduct market (e.g., in gas and in electricity) or in the same product market but in anothergeographic area, the merging firm and the buyer will have more possibilities t ocommunicate with each other and to retaliate if one party defects. The risk of collectivedominance is especially strong in electricity, where the list of conditions conducive t ocollective dominance maps almost exactly on to the characteristics of wholesale electricitymarkets (D. Newbery, 2005).

Thirdly, remedies can also go wrong because of uncertainty. An unexpected event (e.g., a


14

technical or a demand change) can make the remedies ineffective (for instance, ifdivestment remedies in power generation are targeted on base-load plants, whereasunexpected demand growth in the future would have required divestments of flexible peak-load plants). According to a FTC study (1999) based on 35 merger cases completed between1990 and 1994, 25 % of divestitures have not been successful in re-establishing the pre-merger level of competition. The problem is probably worse in markets recently opened t ocompetition like electricity and gas markets, because competition authorities have to makeguesses not only on the effects of the merger and on the effects of the remedies, but also onfuture modes of competition and market boundaries before they are stabilised. By contrastwith other industries, merger control in recently liberalised markets requires competitionauthorities to read in the crystal ball three times. As a result, the risk of errors is higher.

Again, in so far as efficiency is difficult to achieve in mergers related to the electricitysector, whereas the cost of anticompetitive effects can be massive for consumers, one canexpect the costs for consumers of weak type II errors to be higher than the costs of weaktype I errors.

As a conclusion, in the electricity sector the use of merger remedies is a risky game to playwith. To protect consumers, economics recommends merger controls to be more cautiousand stringent than in other sectors, that is to take the risk of being wrong in prohibiting amerger or imposing severe pro-competitive remedies rather than to take the risk of beingwrong in clearing a merger with possibly ineffective remedies.


15

3. Anticompetitive effects of gas-electricity mergers

Gas and electricity are both substitutes and complements. They are substitutes on the finalmarket for certain energy uses (e.g., central heating, water heating, cooking) wheneverconsumers have to change their equipment. They are complements when power is producedby gas turbines and CCGTs. In fact in Spain, as in other countries such as the UK, thegrowth of demand for natural gas is fuelled by power generation. Note also that gas andelectricity supply to final consumers may present redundant costs (e.g., billing costs, searchcosts) that may be eliminated in tying the two energies in a dual fuel offer.

There are therefore good reasons for so-called gas-electricity convergence. It results in thecreation of firms owning assets, exercising capabilities and serving clients in both energies.Such an outcome may be beneficial to consumers depending on whether or not it increasesthe competitive pressures (e.g., a gas company investing in a CCGT plant, an electricitycompany attracting gas customers with a dual fuel offer and price rebates). Whencontrolling a convergent merger, the competition authorities must assess whether theanticompetitive effects offset the benefits of convergence for the consumers.

Four main anticompetitive effects

The main potential anticompetitive effects of gas-electricity M&As are fourfold:

1. A horizontal effect on the electricity wholesale market through the elimination of apotential or effective competitor (i.e., the gas company’s generation activity)

2. A horizontal effect on the electricity and gas retail markets through the elimination ofa potential or effective competitor (i.e., the gas company in electricity retail and theelectricity company in gas retail)

3. A vertical effect on the electricity wholesale market through an input foreclosure (e.g.,higher price of gas to rival power generators)

4. A vertical effect on the gas wholesale market through a customer foreclosure (e.g.,lower demand for rival gas suppliers).

Those multiple anticompetitive effects of convergent mergers are severe enough to haverecently led two leading antitrust authorities to make a rare decision of prohibition. TheBundeskartellamt has disapproved the Eon/Ruhrgas transaction in January 2002 while theEuropean Commission blocked the EdP/GdP transaction in December 2004. These twocases provide a sound basis to illustrate the modus operandi and the dangers of the fouranticompetitive effects mentioned above.

The input and customer foreclosures were pinpointed in both convergent mergers. The


16

general principle of foreclosure consists in raising prices and damaging competitors’performances by restricting access to a key input or a key purchaser. (i) Input foreclosurein a gas-electricity merger: the key input candidate is the gas produced by the upstreamcompany, say Ruhrgas or GdP, and used by the rivals of the downstream company, say thecompetitors of Eon or EdP, who generate power with gas. Ruhrgas and GdP were found ableto control the access of electricity producers to gas supply because of their dominance ingas import and in the wholesale gas market, and their involvement in gas transportation.(ii) Customer foreclosure in a gas-electricity merger: the key purchaser is the electricalcompany of the new entity. Its sourcing of gas for power plants and final dual fuelconsumers may be ensured by the upstream gas division and therefore may be withdrawnfrom the market. In Eon/Ruhrgas, the Bundeskartellamt feared that the merger wouldforeclose as much as 20% of annual gas sales in Germany. In EdP/GdP, the Commissionfeared that, when the market opens to competition, the merger will foreclose all freecompetitive gas demand stemming from EdP for operating its CCGT generation plants andan affiliated local gas distributor in charge of retail supply.

In addition to vertical effects, the decision of European Commission in EdP/GdP alsoillustrates the anticompetitive horizontal effects a gas-electricity merger may raise.Commonly, horizontal effects occur because the two merger firms are effectivecompetitors in certain markets. For instance, GN effectively competes with Endesa on theSpanish electricity wholesale market because GN is present on that market with a marketshare around 5%. Horizontal effects may also occur because firms involved in thetransaction are potential competitors in some markets. In EdP/GdP, the EuropeanCommission found GdP to be a potential competitor of EdP in power generation, wholesalesupply and ancillary services. The Commission established that GdP has strong incentivesand the ability to build new CCGTs plants thanks to (i) its direct, flexible and economicalaccess to gas, (ii) its concern to compensate future market losses when Portuguese gasmarkets open to competition and (iii) its interest in hedging electricity price volatility tocompete on supply in the retail electricity market. As suggested by this last point, theCommission also found that GdP is a potential competitor of EdP vis-à-vis final consumers.Absent the merger, GdP is likely to sell electricity to its gas customers. Reciprocally, EdP isalso a potential competitor of GdP in the retail market. It has incentives to trade gas at alarger scale, to spread its risks and achieve cost savings though economies of scale and t oexpand its activities in gas supply to its electricity customers. Both would enjoy theadvantage of being the incumbent gas or electricity retailer and distributor (large customerbase, well known brand, frequent customer contacts). As well documented by the UK case,such cross entries are the main competitive drive in the retail markets for small customers.

To conclude, it is important to note that vertical and horizontal anticompetitive effects ofconvergent mergers are several, and that they may also be very strong separately:


17

Eon/Ruhrgas was blocked on the grounds of vertical effects alone whereas in the case ofEdP/GdP horizontal effects were claimed to be harmful enough although the two Portuguesefirms were only potential competitors.

The case of the projected GN/Endesa merger

In the case of GN and Endesa, the two merging firms are already today effectivecompetitors. Moreover, GN is also effectively or potentially an important supplier of gasto Endesa and to the other electricity producers for their CCGTs (hereafter, Elcomps). Theconsequence is that none of the four anticompetitive effects mentioned above can beexcluded in case of a merger between GN and Endesa (herefafter, the new entity is notedGNEnd).

They are illustrated in the diagram below. The horizontal arrows represent sources ofcompetition before the merger. Three of the four anticompetitive effects, 1) eliminationof a competitor, i.e. the gas company, in power generation, 2) removal of dual fuelcompetition between the merged companies in retail markets, and 4) customer foreclosurein the wholesale gas market, would simply terminate one source of competition, but effect3) gas input foreclosure in power generation could confer additional market power on themerged company and could allow it to profitably raise the price of gas competinggenerators.

End-Use Energy Customers

GN Other Gas

GasGN

Gas “Coal”Endesa

Gas “Coal”Elcomp

ElectricityGenerationp

1

34

2

Each potential effect therefore deserves a close examination by antitrust authorities.Hereafter, we consider them successively.


18

Horizontal effect on the wholesale electricity market

GN has been a participant in the Spanish wholesale electricity market since 2001. I toperates its own power plants and intends to build more CCGT plants in the near future. I tis therefore already an effective competitor that will be eliminated. This elimination mayraise unilateral and coordinated effects.

The transaction is designed in such a way that (i) the market share of the new entityGNEnd will be below the current market share of Endesa (ii) the market share of Iberdrola,the planned buyer of divested assets, will increase and (iii) the energy mix of Endesa andIberdrola will become more similar.

It is important to understand that such a change does not necessarily mean that only themarket power of Iberdrola may increase. In electricity the potential for market powerdepends whether for a given time (especially peak and off-peak hours) and a given place(especially in load pockets) a generator is necessary in serving demand. It is better measuredby simulation models or, at least, by specific structural indices (e.g., Pivotal Supplier Index,Residual Supply Index) than by market shares in total output. In other terms, the unilateraleffects of a merger greatly depend on the reallocation of plants i.e.g, the plants that areacquired through the merger, the plants that remain within the new entity and the plantsthat are sold to third parties. Only an in-depth investigation may permit us to concludewhether the merger (i) increases the market power only of Iberdrola or of Endesa, (ii)increases the market power of both Endesa and Iberdrola (iii) does not increase the marketpower of either Iberdrola or Endesa.

The current and future concentration in Spanish power generation deserves a few remarks.Note that the pre-merger HHI is already high, around 2700, and could grow higher with theproposed divestments to Iberdrola. In principle, divestitures should decrease the marketshare of the larger supplier and increase the share of a smaller supplier. Then with all elseequal, market power should decrease. For example if shares change from 60/40 to 50/50,market power decreases (with an HHI decrease from 5200 to 5000). However, in thepresent case, the elimination of GN’s small market share and the divesture to an alreadylarge supplier are likely to undo this effect. For example, if shares are (5,40,30,15,10) forGN, Endesa, Iberdrola, and the two other competitors, and GN’s 5% plus 5% from Endesaare transferred to Iberdrola to achieve a distribution of (0,35,40,15,10) market powerincreases (with an HHI increase from 2850 to 3150). Moreover, the merged company couldbe just as happy to give market power to an existing competitor as to itself. In the presentcase, it would be just as important to watch for new market power for Iberdrola as forGNEnd.

Furthermore, the risk of collusion may also increase because of the proposed merger.Certainly, GN is a small player in electricity. It is however a very different player: it is a


19

recent entrant; it does not possess compensation for stranded costs (i.e., CTCs); it onlyowns CCGTs; it is vertically integrated on gas supply. Being a maverick, it lessens the riskof collusion and therefore its elimination increases the risk of collusion. Moreover, asmentioned above, the transaction results in more symmetry between the major Spanishelectricity producers in terms of their energy mixes. Such an increase in symmetry may alsoincrease the risk of collusion: when firms are homogeneous incentives to collude are higher,tacitly agreeing on a common plan of action is easier as well as the monitoring ofdefection. Finally, as pointed out in section 2, the bargaining over remedies betweenincumbents may facilitate collusive practices. Considering all these factors, the proposedtransaction may increase the risk of collusion. An in-depth investigation is necessary t oappreciate it.

In conclusion, the proposed GNEnd merger would lead to increased concentration andmarket power problems in the Spanish electricity wholesale market. Note thatconcentration is already a problem in the Spanish market and therefore it should not beaggravated. For instance, the White Paper of our esteemed colleague I. Perez Arriagaproposes a mechanism to mitigate market power in setting a cap on effective generationcapacity at 22% on-peak and 19% off-peak, that is an HHI around 2200.

Horizontal effect on retail markets

Historically, electricity retail and distribution are ensured in Spain by regional verticallyintegrated monopolies. Today both activities are legally unbundled and the liberalisedmarket has developed. However, Endesa and Iberdrola are still the main players indistribution - they own 80% of the distribution network – and in retail. In their respectivedistribution regions they account for more than 3/4 of the number of connections (orpoints of supply) and of delivered energy. In the still growing gas retail market, GN owns analmost monopoly in distribution and retail.

Unsurprisingly, at a local level, GN is the main challenger of Endesa in the electricity retailmarkets where Endesa is the incumbent electricity distributor and retailer (and reciprocallyin the gas retail markets): GN is the second supplier in terms of points of supply anddelivered energy in Endesa’s distribution regions. As already argued the gas company isusually the most aggressive competitor because it already has a local customer base (its gasconsumers), benefits from an established reputation and can save money in bundlingelectricity with gas. In Catalonia, for instance, 4/5 of households who shifted from Endesato another supplier have contracted with GN. The acquisition of Endesa by GN wouldtherefore eliminate a key competitor. The remaining competition in the local retailmarkets dominated by the merged company would mainly depend on the aggressiveness ofother electricity operators, but without the advantage of being local incumbent suppliers.

In regional markets dominated by other incumbent electricity companies, GNEnd would


20

keep the position of incumbent gas retailer (if the area is supplied in gas), but with anincreased ability to bundle electricity with gas. Yet, if the proposed divestment of gas retailactivities to Iberdrola occurred in the same area where it is already the incumbent electricitydistributor and retailer, the deal would spread the effect of removing a key competitorbeyond the regions where GN and Endesa’s retail activities overlap. At a national level, therisk of an increase in collusion would not be null and would deserve a close analysis.

Vertical effect on the wholesale electricity market

Vertical anticompetitive effects are more complex to analyse than horizontal effects. Theydepend especially on the presence of market power at the upstream or downstream level,on relative performances of competitors, and on the effect of their elimination onconsumers’ surplus. In the following part of this section, we assume that, due to limitedalternative gas sources, GN can influence the supply of gas as a primary fuel to its ownelectricity generators, to the gas-fired plants of Endesa and to the gas-fired plants ofElcomps. Under this assumption, one can expect that the merger would change the verticalrelationship between GN, as an input supplier, and the electricity producers. Whereas theother anticompetitive effects would simply terminate one source of competition, thevertical integration of GN and Endesa would confer additional market power on the mergedcompany and allow it to profitably raise the price of gas competing generators. It isimportant to trace the source of this additional market power through the electricitymarket because it is more inclusive than might be expected.

In the short run, one can expect some efficiency gains for the supply of gas to Endesa’splants. The internal price of supply between two production units belonging to the samecorporate entity is likely to be lower than the price negotiated by a firm with independentfor-profit firms. Consequently, after the merger, the merged entity GNEnd might benefitfrom the same lower price of gas as the GN’s plants uses to pay, lower than the prices paidby Elcomps. This would be beneficial since the reduction of any margin has the positiveeffect of increasing production and reducing the deadweight loss on the intermediary marketfor gas. But it also means that one competitor on the electricity market, GNEnd, wouldhave a lower cost of production. Without strong competitive pressures in power generation,one can fear that GNEnd’s efficiency gains would not be passed on to consumers throughlower prices, but could contribute to harmful horizontal anticompetitive effects (e.g. anincrease in market concentration, as lower production costs should increase the marketshare of GNEnd at the expense of its competitors).

Moreover, the vertical effect would confer additional market power on the mergedcompany and would allow it to profitably raise the price of gas competing generators. GNhas two sources of profit from an increase of the price of gas sold to Elcomps: the revenuefrom the gas sale and the increase in electricity prices paid to its own gas-fired generation.


21

Absent the merger, GN owns relatively little generation; therefore, the electricity priceeffect is relatively weak. After the merger, GN would profit from increased electricityprices not just for its previously-owned generation and the gas-fired generation of Endesa,but it would also profit from increased electricity prices paid to almost all non-gas-firedgeneration owned by Endesa. This is because other types of generation, such as coal, aregenerally selling power whenever gas-fired generation is in use. Raising the market price ofelectricity benefits all producing generators equally. There is always a cost as well as abenefit to exercising market power, and that cost would be a reduction in demand for GN’sgas by Elcomps. But, after a merger, there would be a much greater increase in profits thatcould be used to offset this loss. Consequently greater price markups to Elcomps should beexpected. It might still be argued that end-users could partially protect themselves fromsuch electricity price increases by simply signing long-term contracts with non-gas-firedgenerators (if any) such as coal and nuclear units. But forward contract prices are setrelative to expected spot prices, which feel the full impact of gas-price increases. Henceelectricity customers would gain no protection by using forward markets. Raising the cost ofgas would raise the cost of all electricity produced with lower marginal costs than costs ofgas-fired units. Specifically, if the marginal unit was a gas unit half of the time, and these

units had a cost increase of $2/MWh (or are owned by GN), the average price of electricityover all hours from nuclear, coal, wind and run-of-river hydro would increase by $1/MWh.

This would be true even if only 20% of all power was generated with gas-fired units.

In effect, GNEnd’s default position, which is how much it would earn in the case of anegotiation breakdown, would be enhanced by the takeover as it could use more gas in itsown turbines. If in the short run, the gas price billed to Elcomps is probably fixed because itis part of contracts signed in the past, for the future, harsher negotiations should result in aprogressive increase in the gas price billed to competitors. For the same reason, the mergerwould increase GN’s incentives to keep inefficient allocation mechanisms, such as “firstcome, first served”, for the use of some essential gas infrastructures (pipelines, LNGterminals) in order to keep its market power on gas imports.

But, there is also a risk of contingent foreclosure in peak-load generation. Gas is not anessential input to produce electricity when coal-fired stations are available. However, t oproduce electricity at peak periods, gas is the best choice because it has lower fixed costsand is used in flexible plants that can respond within minutes to the dispatcher’s order.After the hydro reservoirs, gas turbines are the best candidates to do so. When the hydroresources are limited by natural inflows, the largest gas supplier can easily exert a strongpower and exclude some competitors.

Albeit complicated, the vertical effects of the GN/Endesa transaction on the wholesaleelectricity market deserve a close examination.


22

Vertical effect on the wholesale gas market

Today and tomorrow Endesa’s sourcing of gas for power plants and final dual fuelconsumers may be ensured by GN and therefore may be withdrawn from the market. Insubstituting exchanges through market by exchanges within a single integrated firm, themerger may impede the development of the wholesale market, a key aspect of gas andelectricity liberalisation as we already pointed out. Note that the elimination of a keypurchaser of gas may decrease economies of scale that alternative suppliers would haveachieved absent the merger, and then may result in a price increase of their sales. If theintegrated firm has enough market power on the wholesale gas market, it will follow theprice increase rather than offer a lower price.

However, long term contracts for gas concluded before the merger would not be affected inthe short run. The foreclosure effect may concern only a small demand of gas (e.g. CCGTsshort run requirements above long term contracted volumes), but this would still contradictthe objective of an integrated competitive gas market, that is supposed to achieve shortercontract duration and more liquid wholesale markets. Due to the lack of transparency of thewholesale gas markets, more information is needed to assess the possible effect of gasdemand foreclosure. An in-depth investigation is necessary to appreciate it.


23

References

Anderson J., ‘Making operational sense of mergers and acquisitions’ (1999), The ElectricityJournal, Vol. 12, n° 7, August-September, pp 49-59.

Becker-Blease J. R., Goldberg L. G., and Kaen F. R. (2004), ‘Post DeregulationRestructuring of the Electric Power Industry : Value Creation or Value Destruction?’,F e b r u a r y 2 6 , 2 0 0 4 . ( a v a i l a b l e a t :http://papers.ssrn.com/sol3/papers.cfm?abstract_id=625083).

Codognet M-K., J.M.Glachant, F. Lévêque and M-A. Plagnet (2003), "Mergers andAcquisitions in the European Electricity Sector - Cases and patterns", Cerna, 2003.(available at :http://www.cerna.ensmp.fr/Documents/FL-MA-MAsEU-Cases-2003.pdf).

Federal Trade Commission, Bureau of Competition, 1999 ‘A Study of the Commission’sD i v e s t i t u r e P r o c e s s ’ , A u g u s t 6 ( a v a i l a b l e at :http://www.ftc.gov/os/1999/08/index.htm)

Fernando Yañez J. A. (2004), ‘Interactions réglementaires dans les industries énergétiquesen réseaux’, Thèse de Doctorat, Paris I, 24 Juin.

Glachant J.-M. and Lévêque F. (2005), ‘Electricity Internal Market in the EuropeanUnion : What to do next ?’, SESSA report, September 2005 (available athttp://www.sessa.eu.com/public/publications.php)

Hernandez Alva, C.A. (2005) "Competing with your provider: an economic analysis ofmergers in the natural gas and power generation sectors", March, Gremaq, Toulouse( a v a i l a b l e a t :www.fep.up.pt/conferences/earie2005/cd_rom/Session%20V/V.G/alva.pdf

Hunger,D. (2003) ‘Analyzing Gas and Electricity convergence: A supply curve is worth athousand words’ Journal of Regulatory Economics; 21:2 213-222

Motta M. (2004), Competition Policy and Practice, Cambridge University Press

Motta M., Polo M. and Vasconcelos H. (2003) ‘Merger Remedies in the European Union:an Overview’, in Merger Remedies in American and European Union CompetitionLaw, F. Lévêque and H. Shelanski Editors, Edward Elgar, pp 106-128.

Newbery D. (2005), ‘The relationship between regulation and competition policy fornetwork utilities’, paper presented at the Conference in tribute to Jean-JacquesLaffont in Toulouse on 30 June.

US Department of Energy (2000), ‘The Changing Structure of the Electric Power Industry:A n u p d a t e . O c t o b e r 2 ( a v a i l a b l e a t : .http://www.eia.doe.gov/cneaf/electricity/chg_stru_update/update2000.html)

Rey, P. and J. Tirole (2003) ‘A primer on Vertical Integration and Forclosure’. Handbookof Industrial Organization. Edited by Mark Armstrong and Rob Porter. Vol III.

S. Stoft (2002) ‘Power system Economics: Designing Markets for Electricity’, IEEE Press.


24

Appendix : Short bios of signatories to the brief

Julian Barquin, Pontificia Comillas University

Julian Barquin is professor at Universidad Pontificia Comillas de Madrid (ICAI). He is aspecialist of Power Systems with a special interest in the Economics of Energy (he has beenVisiting Scholar at the Department of Applied Economics at the University of Cambridge).He has been involved in the development of different models for the simulation ofoligopolistic and non-oligopolistic electricity markets, including the antecessor of the oneused in the White Paper on Electricity Reform in Spain, as well as in several studies onmarket power assesment in electricity markets.

Lars Bergman, Stockholm School of Economics

Lars Bergman is professor of economics at SSE since 1984. Since 1991 he has an endowedchair in energy and environmental economics. He is a Member of the Royal SwedishAcademy of Engineering Sciences and Chairman of the Swedish Association for EnergyEconomics which serves as a platform for regular meetings between academics, powerindustry representative, regulators and other parties interested in the development of theenergy sector in general and the electricity market in particular. He has served as an expertin several government commissions on energy and environmental policy issues and as anadvisor to the Sydkraft Research Foundation. He is also associated with SNS (Swedish Centrefor Business and Policy Studies) and participates frequently in SNS meetings andconferences with industry representatives.

Claude Crampes, University of Toulouse

Claude Crampes is Professor of economics at University of Toulouse and Research Directorat IDEI (Institut d’Economie Industrielle). He is a specialist of Industrial Organization witha special interest in the Economics of Energy. He published breakthrough articles inEconomica, European Economic Review, Rand Journal of Economics and Journal ofRegulatory Economics. He recently co-signed a paper with N. Fabra on the SpanishElectricity System in a special issue of The Energy Journal edited by D. Newbery.

Jean-Michel Glachant, University Paris XI

Jean-Michel Glachant is permanent professor and Head of the Department of Economics atthe University of Paris XI, as well as head of the “Electricity Group Jean Monnet” at theADIS research center. He edited “Competition in European Electricity Markets: A Cross-Country Comparison” (Edward Elgar, 2003), “The Economics of Contracts. Theory andApplications” (Cambridge University Press, 2002). He took part in the study “Indicators of


25

Internal Market of Electricity” for the European Commission - DG TREN (2000-2001). Herecently coordinated with François Lévêque a large European research programme onElectricity Reforms (SESSA) and co-authored “Electricity internal market in the EuropeanUnion – What to do next?. Jean-Michel Glachant is a member of the International Societyfor New Institutional Economics, member of the International Association for EnergyEconomics, and the Association Française de Science Economique.

Richard Green, University of Birmingham

Richard Green is professor of economics at the University of Birmingham and Director ofthe Institute for Energy Research and Policy. He has been studying the economics andregulation of the electricity industry since 1989, just before the industry in England andWales was privatised. With David Newbery, he was responsible for the most influentialstudy of competition in the British electricity spot market. He has spent a year onsecondment to the Office of Electricity Regulation, and has been a visiting Fellow at theWorld Bank Institute, the University of California Energy Institute and the MasschusettsIntitute of Technology.

Christian Von Hirschhausen, Dresden University of Technology

Christian von Hirschhausen is professor of energy economics and public sector managementat Dresden University of Technology, and research professor at DIW Berlin, the GermanInstitute for Economic Research. He is specializing in applied industrial organization issuesin the energy sector, with a focus on natural gas economics. Previous positions as VisitingProfessor, TU Berlin University of Technology, Workgroup for InfrastructureInfrastructure Policy (WIP), Member of the German Group of Advisors to the Governmentof Ukraine.

François Lévêque, Ecole des mines de Paris

François Lévêque is professor of economics at Ecole des mines de Paris and visitingprofessor at University of California at Berkeley. He is Director at Cerna, the researchcentre of the Ecole des mines in industrial economics. François Lévêque has publishedseveral books in antitrust economics (Antitrust, Patents and Copyright, Edward Elgar, 2005;Merger Remedies in American and European Union Competition Law, Edward Elgar 2003)and in Economics of Regulation (Economie de la réglementation, Editions La Découverte,1999 et 2005; Transport Pricing of Electricity Networks, Kluwer Academic Publishers,2003). He recently coordinated with Jean-Michel Glachant a large European researchprogramme on Electricity Reforms (SESSA) and co-authored with him a policy report“Electricity internal market in the European Union – What to do next?”.

Steven Stoft, Berkeley

Steven Stoft has a twelve-year experience in power market analysis and design. He is the


26

author of Power System Economics, Designing Market for electricity, IEEE, 2002. He isnow an advisor of PJM and California Electricity Oversight Board. Previously. He was aSenior Research Fellow at the University of California Energy Institute, had worked onregulatory and restructuring issues at Lawrence Berkeley National Laboratory and at UCEIand as an economist with the Office of Economic Policy, Federal Energy RegulatoryCommission.